Contract Review

Semantic search across large document sets to find relevant clauses, commitments, or obligations.

The Problem

Contract review across hundreds of documents requires finding specific clauses, obligations, and commitments. Traditional keyword search fails when different contracts use different language for the same concepts — “termination clause” vs “exit provisions” vs “contract end terms”.

Workflow

1. Configure Document Sources

Point Foxhound at your contract repository:

ingestion:
  sources:
    - type: document
      path: "~/contracts/vendor-agreements"
    - type: document
      path: "~/contracts/employment"
    - type: document
      path: "~/contracts/partnership-agreements"

2. Ingest and Explore

uv run python ingest.py --reset
uv run python explore.py --source-type document

3. Search for Specific Clause Types

mkdir -p evidence
 
uv run python query.py \
  --semantic "termination notice period exit clause" \
  --source-type document \
  --top-k 100 --export-json evidence/termination.json
 
uv run python query.py \
  --semantic "liability indemnity limitation damages" \
  --source-type document \
  --top-k 100 --export-json evidence/liability.json
 
uv run python query.py \
  --semantic "payment terms invoice schedule milestone" \
  --source-type document \
  --top-k 100 --export-json evidence/payment.json
 
uv run python query.py \
  --semantic "confidentiality non-disclosure NDA intellectual property" \
  --source-type document \
  --top-k 100 --export-json evidence/confidentiality.json

4. Review Inline

Preview results before running the full analysis:

# Show body previews
uv run python query.py \
  --semantic "termination notice period" \
  --source-type document --show
 
# Read a specific result in full
uv run python query.py \
  --semantic "termination notice period" \
  --source-type document --read 3

5. Merge and Analyse

uv run python merge.py evidence/*.json --output evidence/merged.json
 
uv run python analyze.py evidence/merged.json \
  --full-pipeline \
  --context "Extract and summarise key contractual obligations, notice periods, liability caps, and payment terms" \
  --model deepseek

6. Export Summary

uv run python export.py analysis_output.md --output contract-summary.md

Tips

Semantic search excels here — it matches “30-day notice period” when you search for “termination clause”
Use --source-type document to exclude emails and notes from contract-focused searches
For large contract sets, use gemini-flash with its 1M token context window
Export to CSV with --export clauses.csv for spreadsheet comparison across contracts

Internal Investigations Research Synthesis